River Flow Forecasting using Fuzzy Inference System

Abstract

The Fuzzy Sets Theory has recently been widely and successfully used in engineering problems with complexity, ambiguity, or lack of enough data. The Fuzzy Inference System (FIS) is among these techniques. The main advantage of this technique over traditional methods is that it works based on IF-THEN rules and appoints the relation between input and output variables accordingly. In this study the monthly discharge, temperature, and rainfall are used in the Fuzzy Inference System context as continuous series in order to forecast the river flow discharge for the next months. The effect of each variable in previous time step is determined on the flow discharge in the upcoming month and the best combination and suitable lag time is obtained.